Re: [fsl-discuss] Re: [Patents] The politics of software over-protection -- moving forward..

[Xavi Drudis Ferran <xdrudis@tinet.org>]

El Sun, Apr 27, 2003 at 12:21:27PM -0400, Russell McOrmond deia:
> 
> On Sun, 27 Apr 2003, Xavi Drudis Ferran wrote:
> 
> > Yes I think software is information and all information is created equal. 
> > You cannot base freedom of expression on what kind of expression are you 
> > free to express.
> 
>   I have to ask one of Greg's questions then:  Is your disagreement with
> the patent system for all subject matter, or the patent system for
> specific subject matter?  If specific subject matter, can you articulate 
> the traits of that subject matter sufficient to be interpreted properly in 
> a court?
> 

Yes. Inmaterial creations should be unpatentable, material creations 
patentable. 
The limit should be the use of forces of nature in a new way to solve a 
probllem. 
Is you just use known devices in a new logical combination (e.g. software) 
you 
should not be able to patent it, because it does not take the same investment 
in laboratories, experiments, assays, prototyping, etc. So there is no reason 
to think we would have more of this innovation by granting monopolies on it. 

For physical goods, the economics are different. 

See some more discussion at

http://mail.gnu.org/archive/html/fsfe-uk/2002-03/msg00034.html

this threat may approach some other I already was in

http://mail.gnu.org/archive/html/fsfe-uk/2002-03/msg00034.html

> 
>   A patent application is a piece of information (printed on paper,
> encoded digitally, etc) which describes a method.  It is a kind of
> expression where a limited monopoly is granted by governments on the use
> of this expression.
>

Not at all. A patent is a description of an invention, say a car engine, 
and the patent covers manufacturing, selling. etc. the invention, 
not using the patent. You could use (if there was enough quality) the 
patent to learn and teach, to inspire you a book, and for a hundreds 
of other uses of the expression of the invention that the patent is. 
But you couldn't use the invention (the car engine). 

The confussion does only appear when subject matter is allowed to include
information goods. Then patent = esxpression = description = information = 
software
and it leads the basic patent bargain ad absurdo.
 
>   If we complain that a given patent application does not describe a
> method, we are discussing patent quality (patents written by lawyers and
> not those skilled in the art), not patentability itself.
>
This is not what I was discussing.
 
>   There are many ways to look at software.  I call it governance software
> in ICT for a reason that I believe applies to ICT.  If the software is
> within an industrial process (software that controls a machine that
> manipulates nature) then I would think of it in a different way.
> 

The important think is not where the software is. When deciding
whether to grant a patent, you should look at what does society gain
and what does it loose. It gains knowledge on a new process or
apparatus. It loses freedom of trade. Is it worth it?. It is if the
monopoly does only affect something you would not have otherwise.

It is reasonable to assume that whenever the knowledge supplied by the
patent requires understanding how to use forces of nature in a novel
way, there has been experimentation needed (to validate any model of
the real world you may be using). This has costs that are not present
in the deduction from axioms of the solution to a problem, that is
what any inmaterial solution brings (from math to software or poetry).

Additionaly the production of the apparatus or application of the 
process will necessarily need material substance that will make it 
have some cost at which you can add patent fees. This is not so 
with inmaterial creations. 

>   There is a huge difference between a fiction novel, sheet-music, a
> patent applications or a marriage license (all just human-readable
> information), so why can there not similarly be differences between
> different types of software (machine readable information) which express
> entirely different types of things in entirely different contexts?
>

There is a huge difference on the use of those information examples. 
There is not a big difference in the nature or properties of those 
examples. You can easily reproduce them, you can give the information 
without losing it, the physical instance where you read the information 
is irrelevant, etc. etc. 

On the contrary I can't photocopy a car engine (or at least I can't drive
the car after replacing the engine by its photocopy). 

A monopoly is not the same as a marriage (no jokes here please), nor a
nice sound from an instrument, nor an engine. But all those realities
are different from the description of them that you find in a sheet of
music, a patent applicaction or a marriage license (what's a marriage
license, anyway?). Descriptions are descritions, and what they
describe is different (and the means to describe them are too, of
course, and a piece of information is necessarily distinguishable from
another expressing something different). But the nature of all those
things is similar.

> 
> > Guess what is the name of the discipline studying possible
> > voting systems, decision making organization and the like?. If I'm not
> > mistaken that's "game theory". So it seems hard to distinguish "ICT
> > governance software" from games.
> 
> 
>   First we need to figure out a way to talk to each other, and then try to
> figure out ways to codify the differences in a law.  It is hard for laws
> to distinguish electronic/computer "game theory" (or the codification of
> the control of visual arts/user-directed motion pictures, etc), "ICT

That's not game theory. Game theory is something like "tic tac toe" is
never won if both players are rational, or you can't have a perfect
democratic system (for a particular definition of it that is not
relevant here), or what kind of auctions will be better for what goals.
It's not about killing aliens (although it may have applications there too). 

> governance software", or software used in manufacturing processes partly
> because those who specialize in these types of software haven't been able
> to articulate it either.
> 

No. It is hard to distinguish between different logical creations
because a logical solution to a problem can always be used to solve
other problems. The difficulty in distinguishing is essential, not 
a problem of categorisation or language.

But the reason not to distinguish them in law is not that it is more
or less difficult, it is that they share common properties that lead
to common economic traits that make the patent system counterproductive
for informational goods. 

>   In this forum I have read largely "all or nothing" scenarios, putting
> each camp into an immovable situation -- each has chicken-little "the sky
> is falling" emotions which mean that there is no possibility of productive
> dialog.
>

Not all or nothing. Some people want all patentable, some people may want 
nothing
(I don't hear much of them) and most people here what all information  not 
patentable
and all inventions patentable. To be an invention you have to bring us new
knowledge on the harnessing of forces of nature. 

And pretending the sky isn't falling won't make stay up there (this is a 
joke).

> 
> 
>   If a process is patentable, and components of that process have 
> specialized hardware replaced with standard hardware running specialized 
> software, is the process still patentable?

No. You don't patent the problem, you patent a solution. Some solutions may 
be 
inventions (an analogical guitar pedal) and some may not (a program for a 
computer with a sound card perfoming the same transformations). 

This is not a problem. It just means we have some invention (in the sound 
card)
which allows us to solve a range of problems (many transformations of sound). 

In other words, don't look at the problem you are solving, look at
what are you teaching us. If you use a sound card to produce some
sound that breaks the monitor crystal screen, and it wasn't previously
know that that sound would have that effect, you could patent that,
because you just provided new info on how to use forces of nature to
break that monitor. If it was already known that kind of sound would
break the monitor screen and you just used a general tool to produce
that specific sound with a new program, your contribution is of a logical 
nature and cannot be patented. 

>   The US Supreme court says it is still patentable.
> 

I'm currently interested in what _we_ think that should be patentable, 
regardless of what is patentable here or there.

> 
>   If a board game is not patentable, and components of that game are moved
> to software, is the game still not patentable?

Of course. It was logics, and still is logics. 

> 
>   I happen to think that at this level the decision making is still fine
> (I can live with the Diamond v. Diehr decision, but *not* the State Street
> decision), but this points to either a contradiction (software is both
> patentable and not patentable at the same time - an uncertainty in law
> where the purpose of law is to create certainty), or a recognition that
> not all software can be considered the same subject matter for the purpose
> of patent law.
>

The fact that software is used in an invention should not make it less
of an invention, the fact that a non-invention is or has software
should not make it an invention. Just as the fact that an invention
has a drawing painted on it, or instructions written on it does not
make it unpatentable, and putting words or drawings in an
non-invention does not make it an invention.  You have to look at what
new teaching on use of forces of nature the patent discloses, not at
how it is expressed (in software, drawings, poetry...).
 
> 
>   What if an algorithm used in an industrial process becomes relevant to a
> game (or the reverse)?  My belief is that when an algorithm is taken out
> of the context of its subject area, that it should be re-evaluated in the
> new context.  Software in a game should never be patentable regardless of
> the origins of the method, including if the method originated within an
> industrial method.  Moving the other way is unlikely given that the game
> would be prior art against filing a patent on the method itself in an
> industrial process, but a more narrow patent for the use of that method
> within that new industrial process probably would be patentable.
> 

If I understand what you say, you are proposing the technical invention 
criteria, just as I have been saying, or at least you show the intuition
whose logical conclusion is what I said. You are saying that you can patent
an industrial process (regardless of whether it uses software) but you 
cannot patent the software it uses, since uses of it should be free. 
Now think why is this your inuition. Because what you consider worth of
patent cannot be in the software, but in the industrial process. Information
can never be patented, uses of forces of nature may.  

> 
> 
>   I believe ICT governance software is entirely different from industrial
> methods including software or games, and has very different issues to deal
> with.  Who is harmed by a monopoly in ICT is entirely different than in an
> industrial process, for all the reasons I've mentioned about communications
> rights, as well as all the reasons often given why "software is different"
> (barriers to market entry, etc -- something that may not apply to software
> in an industrial process).
> 

It can perfectly apply to software in an industrial process, as soon
as you remove the industrial process (for instance think of using that
software for simulation of the industrial process). Software is
software, formulas are formulas, and theorems are theorems, but in
order to have an invention, you have to solve real world problems in a
word that is not accessible to math alone, and you need
experimentation. The fact that software is embedded in a material 
object that has different economics, does not mean that you should 
encourage innovations in embedded software alone with monopolies. 
Because the economics of writing those programs are the same, you 
won't have more embedded software by alowing to patent it (by allowing 
to patent a device whose only difference with prior art is software 
that does not express any new insight on harnessing forces of nature). 
So you don't want to patent it.

> 
>   There is a large number of software creators in this forum.  What domain
> are each of you from?
> 

?. I can't see why you ask. I've programmed database applications, 
web applications, some small search engines. Nothing worth of mention.

>   I may be guessing wrong, but most of the people here seem to be in the
> ICT (Internet/networking, office productivity, etc) or infrastructure
> software (operating systems, etc)  domains.  Are we really discussing what
> should be the case in our specific domain, or are we really talking about
> all domains?
>

I'm talking about how to limit subject matters in a way that may be useful 
to society.  

> 
>   If you are talking about all software, are you also talking about all
> patents?  While it may be hard to draw a line (especially one that will be
> interpreted by some of the extremely self-interested patent
> consultants/lobbiests, patent offices and patent courts) we either need to
> draw a line or decide on an "all or nothing".
> 

We need to draw a line and we need to find neutral people to apply it. 
If the people who apply law are corrupt you won't get good results even
when perfect law. So we both good law and good judges. One can't replace 
the other. 

> 
>   I obviously cannot accept "all" (unlimited patentability), and would be
> entirely unwilling to obey such laws if they did exist.  I already
> deliberately broke the intent of the USA's DMCA here in Canada and
> submitted this fact very publicly to Industry Canada
> <http://www.flora.ca/copyright-2001-cmpda-reply.shtml>.  If Canada were to
> adopt legal protection for TPM or unlimited patentability I would continue
> to very publicly infringe these claimed rights in the ICT field as a form
> of civil disobedience.
> 

Good. So we come to what's the limit?. What's yours?. Unfair question, 
I haven't read your whole paper. But it looked on first sight you 
wanted to allowed unlimited patentability and then work out exceptions 
for free software. I can't see what advantages you think there would 
be for non-free software if it could be patented (apart from failing 
to see this proposal as systematic).

> 
>   On the other hand, I also don't agree with those who want to go
> cold-turkey and say "no patentability".  If society decided to abrogate
> patent law, it is something that needs to happen very gradually given the
> incumbent industrial business models in many fields are entirely dependent
> on these monopolies.  Huge changes to government incentives (accountable
> direct funding of research rather than indirect subsidization,
> protectionism, or monopoly creation) will need to be worked out which will
> take considerable amounts of time.
>

I'd say everybody and his mother knows there are problems with the
patent system. I believe in the area of inmaterial innovations they
cannot be solved, because they are essential problems. In the area of
technical inventions they may be solvable, so no need to throw the
baby with the wash water and eradicate the patent system.  

I can't prove that the patent system can be fixed for technical
inventions or that it cannot not, or that it can be fixed but the
price is too high.  I don't know. So I can't argue for eradication of
the patent system, only for limiting it to areas were it may be useful
and then trying to make it useful there.

But people arguing for unlimited patentability may be muddling the
wash water so much that one day we won't be able to see the child any
more, and after n unsuccessful attempts to fix a fundamentally flawed
patent system, people may decide to eradicate it. That might be worse
than restricting it to technical inventions and fixing it there.


-- 
Xavi Drudis Ferran
xdrudis@tinet.org